Você está na página 1de 34

1

TABLE OF CONTENTS

PROBABILITY THEORY


Lecture 1 Basics
Lecture 2 Independence and Bernoulli Trials
Lecture 3 Random Variables
Lecture 4 Binomial Random Variable Applications, Conditional
Probability Density Function and Stirlings Formula.
Lecture 5 Function of a Random Variable
Lecture 6 Mean, Variance, Moments and Characteristic Functions
Lecture 7 Two Random Variables
Lecture 8 One Function of Two Random Variables
Lecture 9 Two Functions of Two Random Variables
Lecture 10 Joint Moments and Joint Characteristic Functions
Lecture 11 Conditional Density Functions and Conditional Expected Values
Lecture 12 Principles of Parameter Estimation
Lecture 13 The Weak Law and the Strong Law of Large numbers
2
1. Basics
Probability theory deals with the study of random
phenomena, which under repeated experiments yield
different outcomes that have certain underlying patterns
about them. The notion of an experiment assumes a set of
repeatable conditions that allow any number of identical
repetitions. When an experiment is performed under these
conditions, certain elementary events occur in different
but completely uncertain ways. We can assign nonnegative
number as the probability of the event in various
ways:
), (
i
P
i

PROBABILITY THEORY
PILLAI
3
Laplaces Classical Definition: The Probability of an
event A is defined a-priori without actual experimentation
as


provided all these outcomes are equally likely.
Consider a box with n white and m red balls. In this case,
there are two elementary outcomes: white ball or red ball.
Probability of selecting a white ball
We can use above classical definition to determine the
probability that a given number is divisible by a prime p.
,
outcomes possible of number Total
to favorable outcomes of Number
) (
A
A P =
.
m n
n
+
=
(1-1)
PILLAI
4
If p is a prime number, then every p
th
number (starting
with p) is divisible by p. Thus among p consecutive integers
there is one favorable outcome, and hence



Relative Frequency Definition: The probability of an
event A is defined as


where n
A
is the number of occurrences of A and n is the
total number of trials.
We can use the relative frequency definition to derive
(1-2) as well. To do this we argue that among the integers
the numbers are divisible by p.
{ }
1
P a given number is divisible by a prime p
p
= (1-2)
n
n
A P
A
n
lim ) (

=
(1-3)
PILLAI
1, 2, 3, , , n
2 , , p p
5
Thus there are n/p such numbers between 1 and n. Hence



In a similar manner, it follows that


and



The axiomatic approach to probability, due to Kolmogorov,
developed through a set of axioms (below) is generally
recognized as superior to the above definitions, (1-1) and
(1-3), as it provides a solid foundation for complicated
applications.
{ }
/
1
.

lim
n
n p
n
p
P a given number N is divisible by a prime p

= = (1-4)
{ }
2
2
1
P p divides any given number N
p
=
(1-6)
(1-5)
{ }
1
. P pq divides any given number N
pq
=
PILLAI
6
The totality of all known a priori, constitutes a set O,
the set of all experimental outcomes.


O has subsets Recall that if A is a subset of
O, then implies From A and B, we can
generate other related subsets etc.
,
i

{ } , , , ,
2 1 k
= O
(1-7)
A e . O e
. , , , C B A
, , , , B A B A B A
(1-8)
and
{ }
{ } B A B A
B A B A
e e =
e e =


and |
or |
{ } A A e = |
PILLAI
7
A B
B A
A B A
B A
A
If the empty set, then A and B are
said to be mutually exclusive (M.E).
A partition of O is a collection of mutually exclusive
subsets of O such that their union is O.
, | = B A
. and ,
1
O = =
=

i
i j i
A A A |
B A
| = B A
1
A
2
A
n
A
i
A
A
(1-9)
j
A
Fig. 1.2
Fig.1.1
PILLAI
8
De-Morgans Laws:
B A B A B A B A = = ;
A B
B A
A B
B A
A B
B A
A B
Often it is meaningful to talk about at least some of the
subsets of O as events, for which we must have mechanism
to compute their probabilities.
Example 1.1: Consider the experiment where two coins are
simultaneously tossed. The various elementary events are
Fig.1.3
PILLAI
(1-10)
9
{ }. , , ,
4 3 2 1
= O
) , ( ), , ( ), , ( ), , (
4 3 2 1
T T H T T H H H = = = =
and
The subset is the same as Head
has occurred at least once and qualifies as an event.
Suppose two subsets A and B are both events, then
consider
Does an outcome belong to A or B
Does an outcome belong to A and B
Does an outcome fall outside A?
{ } , ,
3 2 1
= A
B A =
B A =
PILLAI
10
Thus the sets etc., also qualify as
events. We shall formalize this using the notion of a Field.
Field: A collection of subsets of a nonempty set O forms
a field F if



Using (i) - (iii), it is easy to show that etc.,
also belong to F. For example, from (ii) we have
and using (iii) this gives
applying (ii) again we get where we
have used De Morgans theorem in (1-10).
, , , , B A B A B A
. then , and If (iii)
then , If (ii)
(i)
F B A F B F A
F A F A
F
e e e
e e
e O
, , B A B A
, , F B F A e e ; F B A e
, F B A B A e =
PILLAI
(1-11)
11
Thus if then

From here on wards, we shall reserve the term event
only to members of F.
Assuming that the probability of elementary
outcomes of O are apriori defined, how does one
assign probabilities to more complicated events such as
A, B, AB, etc., above?
The three axioms of probability defined below can be
used to achieve that goal.
, , F B F A e e
{ }. , , , , , , , , B A B A B A B A B A F O =
) (
i i
P p =
i

PILLAI
(1-12)
12
Axioms of Probability
For any event A, we assign a number P(A), called the
probability of the event A. This number satisfies the
following three conditions that act the axioms of
probability.



(Note that (iii) states that if A and B are mutually
exclusive (M.E.) events, the probability of their union
is the sum of their probabilities.)
). ( ) ( ) ( then , If (iii)
unity) is set whole the of ty (Probabili 1 ) ( (ii)
number) e nonnegativ a is ty (Probabili 0 ) ( (i)
B P A P B A P B A
P
A P
+ = =
= O
>
|
PILLAI
(1-13)
13
The following conclusions follow from these axioms:
a. Since we have using (ii)

But and using (iii),

b. Similarly, for any A,
Hence it follows that
But and thus
c. Suppose A and B are not mutually exclusive (M.E.)?
How does one compute
, O = A A
. 1 ) ( ) P( = O = P A A
, | = A A
). ( 1 ) or P( 1 ) P( ) ( ) P( A P A A A P A A = = + =
(1-14)
{ } { }. | | = A
{ } ( ) . ) ( ) ( | | P A P A P + =
{ } , A A = | { } . 0 = | P
(1-15)
? ) ( = B A P
PILLAI
14
To compute the above probability, we should re-express
in terms of M.E. sets so that we can make use of
the probability axioms. From Fig.1.4 we have

where A and are clearly M.E. events.
Thus using axiom (1-13-iii)

To compute we can express B as

Thus

since and are M.E. events.
B A
, B A A B A =
(1-16)
). ( ) ( ) ( ) ( B A P A P B A A P B A P + = =
), ( B A P
A B BA A B A B
A A B B B
= =
= O =
) ( ) (
) (
), ( ) ( ) ( A B P BA P B P + =
AB BA= B A A B =
B A
(1-17)
(1-18)
(1-19)
A B A
B A
Fig.1.4
PILLAI
15
From (1-19),

and using (1-20) in (1-17)

Question: Suppose every member of a denumerably
infinite collection A
i
of pair wise disjoint sets is an
event, then what can we say about their union

i.e., suppose all what about A? Does it
belong to F?
Further, if A also belongs to F, what about P(A)?
) ( ) ( ) ( AB P B P B A P =
). ( ) ( ) ( ) ( AB P B P A P B A P + =
?
1

=
=
i
i
A A
, F A
i
e
(1-22)
(1-20)
(1-21)
(1-23)
(1-24)
PILLAI
16
The above questions involving infinite sets can only be
settled using our intuitive experience from plausible
experiments. For example, in a coin tossing experiment,
where the same coin is tossed indefinitely, define
A = head eventually appears.
Is A an event? Our intuitive experience surely tells us that
A is an event. Let


Clearly Moreover the above A is
{ }
} , , , , , {
toss th the on 1st time for the appears head
1
h t t t t
n A
n
n

=
=
(1-26)
. | =
j i
A A
.
3 2 1
=
i
A A A A A
(1-27)
(1-25)
PILLAI
17
We cannot use probability axiom (1-13-iii) to compute
P(A), since the axiom only deals with two (or a finite
number) of M.E. events.
To settle both questions above (1-23)-(1-24), extension of
these notions must be done based on our intuition as new
axioms.
o-Field (Definition):
A field F is a o-field if in addition to the three conditions
in (1-11), we have the following:
For every sequence of pair wise disjoint
events belonging to F, their union also belongs to F, i.e.,
, 1 , = i A
i
.
1
F A A
i
i
e =

=

(1-28)
PILLAI
18
In view of (1-28), we can add yet another axiom to the
set of probability axioms in (1-13).
(iv) If A
i
are pair wise mutually exclusive, then

Returning back to the coin tossing experiment, from
experience we know that if we keep tossing a coin,
eventually, a head must show up, i.e.,

But and using the fourth probability axiom
in (1-29),

). (
1 1

=
=
|
|
.
|

\
|
n
n
n
n
A P A P

(1-29)
. 1 ) ( = A P
(1-30)

=
=
1
,
n
n
A A
). ( ) (
1 1

=
=
|
|
.
|

\
|
=
n
n
n
n
A P A P A P

(1-31)
PILLAI
19
From (1-26), for a fair coin since only one in 2
n
outcomes
is in favor of A
n
, we have


which agrees with (1-30), thus justifying the
reasonableness of the fourth axiom in (1-29).
In summary, the triplet (O, F, P) composed of a nonempty
set O of elementary events, a o-field F of subsets of O, and
a probability measure P on the sets in F subject the four
axioms ((1-13) and (1-29)) form a probability model.
The probability of more complicated events must follow
from this framework by deduction.
, 1
2
1
) ( and
2
1
) (
1 1
= = =


=

= n
n
n
n
n
n
A P A P (1-32)
PILLAI
20
Conditional Probability and Independence
In N independent trials, suppose N
A
, N
B
,

N
AB
denote the
number of times events A, B and AB occur respectively.
According to the frequency interpretation of probability,
for large N

Among the N
A
occurrences of A, only N
AB
of them are also
found among the N
B
occurrences of B. Thus the ratio
. ) ( , ) ( , ) (
N
N
AB P
N
N
B P
N
N
A P
AB B A
~ ~ ~
(1-33)
) (
) (
/
/
B P
AB P
N N
N N
N
N
B
AB
B
AB
= =
(1-34)
PILLAI
21
is a measure of the event A given that B has already
occurred. We denote this conditional probability by
P(A|B) = Probability of the event A given
that B has occurred.
We define


provided As we show below, the above definition
satisfies all probability axioms discussed earlier.
,
) (
) (
) | (
B P
AB P
B A P =
. 0 ) ( = B P
(1-35)
PILLAI
22
We have
(i)

(ii) since O B = B.

(iii) Suppose Then


But hence

satisfying all probability axioms in (1-13). Thus (1-35)
defines a legitimate probability measure.
, 0
0 ) (
0 ) (
) | ( >
>
>
=
B P
AB P
B A P
.
) (
) (
) (
) ) ((
) | (
B P
CB AB P
B P
B C A P
B C A P

=

=
, 1
) (
) (
) (
) (
) | ( = =
O
= O
B P
B P
B P
B P
B P
), | ( ) | (
) (
) (
) (
) (
) | ( B C P B A P
B P
CB P
B P
AB P
B C A P + = + =
| = C A
, | = BC AB
). ( ) ( ) ( CB P AB P CB AB P + =
(1-39)
(1-37)
(1-36)
(1-38)
PILLAI
23
Properties of Conditional Probability:
a. If and

since if then occurrence of B implies automatic
occurrence of the event A. As an example, but

in a dice tossing experiment. Then and
b. If and


, , B AB A B = c
1
) (
) (
) (
) (
) | ( = = =
B P
B P
B P
AB P
B A P
(1-40)
, A B c
). (
) (
) (
) (
) (
) | ( A P
B P
A P
B P
AB P
B A P > = =
, , A AB B A = c
(1-41)
, A B c
. 1 ) | ( = B A P
PILLAI
{outcome is even}, ={outcome is 2}, A B =
24
(In a dice experiment,
so that The statement that B has occurred (outcome
is even) makes the odds for outcome is 2 greater than
without that information).
c. We can use the conditional probability to express the
probability of a complicated event in terms of simpler
related events.

Let are pair wise disjoint and their union is O.
Thus and

Thus
. B Ac
.
1
O =
=

n
i
i
A
n
A A A , , ,
2 1

, | =
j i
A A
. ) (
2 1 2 1 n n
BA BA BA A A A B B = =
(1-42)
(1-43)
PILLAI
{outcome is 2}, ={outcome is even}, A B =
25
But so that from (1-43)


With the notion of conditional probability, next we
introduce the notion of independence of events.
Independence: A and B are said to be independent events,
if

Notice that the above definition is a probabilistic statement,
not a set theoretic notion such as mutually exclusiveness.
). ( ) ( ) ( B P A P AB P = (1-45)
, | | = =
j i j i
BA BA A A

= =
= =
n
i
i i
n
i
i
A P A B P BA P B P
1 1
). ( ) | ( ) ( ) ( (1-44)
PILLAI
26
Suppose A and B are independent, then

Thus if A and B are independent, the event that B has
occurred does not shed any more light into the event A. It
makes no difference to A whether B has occurred or not.
An example will clarify the situation:
Example 1.2: A box contains 6 white and 4 black balls.
Remove two balls at random without replacement. What
is the probability that the first one is white and the second
one is black?
Let W
1
= first ball removed is white
B
2
= second ball removed is black
). (
) (
) ( ) (
) (
) (
) | ( A P
B P
B P A P
B P
AB P
B A P = = =
(1-46)
PILLAI
27
We need We have
Using the conditional probability rule,

But

and

and hence
? ) (
2 1
= B W P
). ( ) | ( ) ( ) (
1 1 2 1 2 2 1
W P W B P W B P B W P = =
,
5
3
10
6
4 6
6
) (
1
= =
+
= W P
,
9
4
4 5
4
) | (
1 2
=
+
= W B P
.
90
24
9
4
10
6
) (
2 1
= = B W P
.
1 2 2 1 2 1
W B B W B W = =
(1-47)
PILLAI
28
Are the events W
1
and B
2
independent? Our common sense
says No. To verify this we need to compute P(B
2
). Of course
the fate of the second ball very much depends on that of the
first ball. The first ball has two options: W
1
= first ball is
white or B
1
= first ball is black. Note that
and Hence W
1
together with B
1
form a partition.
Thus (see (1-42)-(1-44))


and

As expected, the events W
1
and B
2
are dependent.
,
1 1
| = B W
.
1 1
O = B W
,
5
2
15
2 4
5
2
3
1
5
3
9
4
10
4
3 6
3
5
3
4 5
4

) ( ) | ( ) ( ) | ( ) (
1 1 2 1 1 2 2
=
+
= + =
+
+
+
=
+ = B P B B P W P W B P B P
.
81
20
) (
5
3
5
2
) ( ) (
1 2 1 2
= = = W B P W P B P
PILLAI
29
From (1-35),

Similarly, from (1-35)

or

From (1-48)-(1-49), we get

or

Equation (1-50) is known as Bayes theorem.
). ( ) | ( ) ( B P B A P AB P =
,
) (
) (
) (
) (
) | (
A P
AB P
A P
BA P
A B P = =
). ( ) | ( ) ( A P A B P AB P =
(1-48)
(1-49)
). ( ) | ( ) ( ) | ( A P A B P B P B A P =
(1-50)
) (
) (
) | (
) | ( A P
B P
A B P
B A P =
PILLAI
30
Although simple enough, Bayes theorem has an interesting
interpretation: P(A) represents the a-priori probability of the
event A. Suppose B has occurred, and assume that A and B
are not independent. How can this new information be used
to update our knowledge about A? Bayes rule in (1-50)
take into account the new information (B has occurred)
and gives out the a-posteriori probability of A given B.
We can also view the event B as new knowledge obtained
from a fresh experiment. We know something about A as
P(A). The new information is available in terms of B. The
new information should be used to improve our
knowledge/understanding of A. Bayes theorem gives the
exact mechanism for incorporating such new information.
PILLAI
31
A more general version of Bayes theorem involves
partition of O. From (1-50)


where we have made use of (1-44). In (1-51),
represent a set of mutually exclusive events with
associated a-priori probabilities With the
new information B has occurred, the information about
A
i
can be updated by the n conditional probabilities
,
) ( ) | (
) ( ) | (
) (
) ( ) | (
) | (
1

=
= =
n
j
j j
i i i i
i
A P A B P
A P A B P
B P
A P A B P
B A P
(1-51)
, 1 , n i A
i
=
47). - (1 using , 1 ), | ( n i A B P
i
=
. 1 ), ( n i A P
i
=
PILLAI
32
Example 1.3: Two boxes B
1
and

B
2
contain 100 and 200
light bulbs respectively. The first box (B
1
) has 15 defective
bulbs and the second 5. Suppose a box is selected at
random and one bulb is picked out.
(a) What is the probability that it is defective?
Solution: Note that box B
1
has 85 good and 15 defective
bulbs. Similarly box B
2
has 195 good and 5 defective
bulbs. Let D = Defective bulb is picked out.
Then
. 025 . 0
200
5
) | ( , 15 . 0
100
15
) | (
2 1
= = = = B D P B D P
PILLAI
33
Since a box is selected at random, they are equally likely.

Thus B
1
and B
2
form a partition as in (1-43), and using
(1-44) we obtain


Thus, there is about 9% probability that a bulb picked at
random is defective.
.
2
1
) ( ) (
2 1
= = B P B P
. 0875 . 0
2
1
025 . 0
2
1
15 . 0
) ( ) | ( ) ( ) | ( ) (
2 2 1 1
= + =
+ = B P B D P B P B D P D P
PILLAI
34
(b) Suppose we test the bulb and it is found to be defective.
What is the probability that it came from box 1?

Notice that initially then we picked out a box
at random and tested a bulb that turned out to be defective.
Can this information shed some light about the fact that we
might have picked up box 1?
From (1-52), and indeed it is more
likely at this point that we must have chosen box 1 in favor
of box 2. (Recall box1 has six times more defective bulbs
compared to box2).
. 8571 . 0
0875 . 0
2 / 1 15 . 0
) (
) ( ) | (
) | (
1 1
1
=

= =
D P
B P B D P
D B P
? ) | (
1
= D B P
(1-52)
; 5 . 0 ) (
1
= B P
, 5 . 0 857 . 0 ) | (
1
> = D B P
PILLAI